Kelly Criterion Calculator
100% private — runs on your device, never uploaded. Works offline once loaded.
Enter your win probability and your average win-to-loss payoff ratio to calculate the Kelly-optimal fraction of your bankroll to stake, along with the safer half-Kelly and quarter-Kelly stakes. Everything is computed in your browser.
The Kelly Criterion in one formula
The Kelly Criterion, introduced by John L. Kelly Jr. in 1956, answers a deceptively simple question: given a bet with a known edge, what fraction of your bankroll should you stake to grow your money as fast as possible over the long run? Stake too little and you leave growth on the table; stake too much and a run of losses can wipe you out.
For a bet that pays R to 1 (your average win divided by your average loss) and wins with probability W, the optimal fraction is f = W − (1 − W) / R. This tool takes W as a percentage and R as a plain ratio, then reports f as a percentage of your bankroll.
Why full Kelly is rarely used raw
Full Kelly maximises the expected logarithm of wealth, which is the mathematically fastest path to growth — but it is also punishingly volatile. Drawdowns of 50% or more are entirely normal at full Kelly, and the formula assumes your estimate of W and R is exactly correct. In reality those inputs are noisy guesses.
Because overbetting is far more dangerous than underbetting, practitioners commonly scale the stake down. Half Kelly captures roughly three-quarters of the growth rate of full Kelly with only half the volatility, which is why it is a popular default. This tool shows full, half and quarter Kelly side by side so you can choose your comfort level.
- Full Kelly: fastest theoretical growth, highest volatility
- Half Kelly: ~75% of the growth, ~50% of the variance
- Quarter Kelly: very conservative, smooth equity curve
- Overbetting beyond full Kelly reduces growth AND raises risk
When it applies — and when it doesn't
Kelly assumes you can estimate your edge honestly, that outcomes are independent, and that you can reinvest continuously. It works best for repeated bets with a genuine, measurable advantage: sports betting models, options and trading strategies with a backtested edge, or any situation with a positive expected value.
It breaks down when your probability estimate is unreliable, when a single loss can be catastrophic, or when bets are correlated so that many positions lose at once. Treat the output as an upper bound on a sensible stake, not a mandate — and never stake more than you can afford to see cut in half.
Frequently asked questions
How do I calculate the payoff ratio R?
Divide your average winning amount by your average losing amount. If wins average $200 and losses average $100, R is 2. For an even-money bet, R is 1.
Is win probability the same as odds?
No. Win probability is your estimated chance of winning as a percentage (e.g. 55%). The payoff ratio R captures the odds — how much you win relative to how much you lose.
What if my win probability is above 50% but Kelly is still low?
A high win rate with small wins and large losses can still produce a small or negative Kelly fraction. Both the probability and the payoff size matter; the formula weighs them together.
Can the Kelly fraction be more than 100%?
Yes, when you have a large edge with favourable odds. In practice this would require leverage, which reintroduces ruin risk, so most people cap the stake well below 100%.
Does Kelly guarantee I won't lose money?
No. It only maximises long-run growth for a genuine edge. If your estimated edge is wrong or non-existent, Kelly cannot save you — and short-term drawdowns are expected even when it is correct.
Is my data private?
Yes. The calculation runs entirely in your browser; none of the numbers you enter are uploaded or stored anywhere.
Advertisement